Utilising small data and technology to drive manufacturing improvement

Author: Steve Wilkinson, Chief Technical Officer at Cimlogic

There is a lot of talk about big data these days – “data is the new oil” and “the world’s most valuable resource is no longer oil, but data”. There is an incredible amount of data available in the digital world which will only get bigger in the future. What are manufacturers, large and small going to do with all this data and how can they gain real value from it?

Big data is about machinery and collecting large data sets, whereas small data is about people. Generally, people can be very unpredictable whilst machines are usually predictable, if they are programmed and maintained by people properly.  We find that most production issues are related to people and processes and not machines.

The DIKW Pyramid Model

Then DIKW model is a great place to start when embarking on a continuous improvement journey. At the bottom of the pyramid, information is extracted from the data which is turned into knowledge and then transformed into Wisdom. Wisdom is discovering what you know now that you didn’t know before, to make informed decisions – it’s only at this point that value can be gained from the data.

For years we have seen a disconnect between the data and wisdom layer within manufacturing. Decisions are often made by people’s gut feeling and experience, without real facts or data wisdom. The goal here is to realign data and wisdom, so that decisions are made on what the data is telling us.

Many of our larger clients on digital journeys connect everything they can to a network server to get data. This bottom up approach is very costly, time consuming and will not see a quick return. According to McKinsey research most companies are only analysing 1% of their data set. This raises the question; why are manufacturers logging all this data in a bottom-up approach, when most of it goes to waste anyway?

The traditional approach to transforming data

Transforming data to wisdom has been a very slow process in the past, with technology layers connecting between numerous automation devices, control systems (PLC, SCADA), ERP and MES systems. The new technologies available now such as Industrial Internet of Things (IIoT) devices can speed up the process considerably. However, the drivers to getting the data and making better decisions haven’t changed; the way to do it has improved somewhat and become much quicker.

Start off small

Start with something that can demonstrate value and decide what to concentrate your efforts on, forgetting about the huge amounts of dark data that wouldn’t be used anyway. Don’t embark on a project for technology’s sake, by choosing the technology first and then logging the data. Think about what you want to know, the value of getting it and then work back to the data you need.

Real life solutions

A continuous packaging line with a bottle filling process could get quick wins from a single sensor on the production line monitoring line speed, short stops and changeovers. An operator may capture downtime data on paper or excel in intervals of 5 minutes. As a line operator there is no ability to visualise the production line speed to judge whether it is running at 450 bottles per minute or 470 bottles per minute for example. This just cannot be achieved in real-time! However, a single sensor on the line can give real-time speed and expose short stops of less than 5 minutes, that won’t be recorded manually.

A changing landscape

Technology is considerably cheaper nowadays, short cutting the traditional approach in a much quicker and easier way. Shop floor data can instantly reach the decision makers using technologies such as IIoT, mobile devices and cloud computing. Smart devices and wireless connectivity remove the need for large, costly infrastructure projects. This enables manufacturers to access the data to justify acting on the low hanging fruit opportunities.

Small data is all about people, so it’s important that everyone involved in the improvement process buys into the technology and trusts what the data is telling them. Very little investment and time is needed to gain tangible benefits from small data. When manufacturers see real value from the data, they are more likely to reinvest in another improvement initiative, starting the knowledge pyramid process over again.

About Steve Wilkinson – CTO, Cimlogic

Steve Wilkinson is the Chief Technology Officer for Cimlogic, empowering manufacturers to become World Class Manufacturers, by providing them with the technology and expertise to drive manufacturing productivity.

Steve first began his career with Cimlogic 15 years ago as a Control Systems Engineer. In 2012, he moved into the role of Solutions Architect. His role within Cimlogic involves setting the technological roadmap for Cimlogic and its clients from sectors within Food and Beverage, Pharmaceutical, Consumer Packaged Goods, Petrochemical and Animal Nutrition.